3.2.1. Introduction

Population projections are arguably the backbone of GHG emissions scenarios, and are comparable in some ways with them. Population projections cover timeframes of a century or more, and they involve social and economic considerations and uncertainties similar to those in GHG emissions scenarios. Population projections are among the most commonly cited indicators of the future state of the world. Compared to the multitude of projection efforts they have a relatively high accuracy in the near-to-medium term. Even so the future is always unknowable and surprises are in store, as confirmed by a cursory review of the past history of population projections in which fundamental events were largely unforeseen (post-World War II baby boom, acquired immunodeficiency syndrome (AIDS) or the recent rapidity of fertility decline in developing countries).

To be useful for the development of emissions scenarios, population projections need a timeframe of a century or more, global coverage and regional disaggregations, and an appropriate treatment of uncertainty reflected in the variants of the projections. Although other "demographic units" more immediately linked to GHG emissions than people, such as automobiles or households, can be considered, the integrated assessment models used in this report are all based on regional population and, in some cases, labor-force projections.

3.2.2. Past Population Trends

World annual population growth rates probably averaged less than 0.6% during the 18th and 19th centuries, passed the 1% rate around 1920, and peaked at 2.04% in the late 1960s (UN, 1998). This peak coincided with growing international concern about population growth in general. World population reached 1 billion in 1804, 2 billion in 1927, 3 billion in 1960, 4 billion in 1974, and 5 billion in 1987, reaching the 6 billion level shortly before the millenium (UN, 1998).

The population of the developing regions increased from 1.71 billion in 1950 to 4.59 billion in 1996, with annual growth rates dropping from a peak of 2.5% in 1965 to 1.7% presently. The population of the more-developed regions increased from 813 million to 1.18 billion over the same period, with annual growth rates dropping from 1.2% in 1950 to 0.4% presently (UN, 1998). Population distribution and growth thus differ markedly among major geographic regions. Latin America and the Caribbean was the fastest growing region between 1950 and 1970, followed by Africa, and this is projected to remain the case until 2050 (UN, 1998). Table 3-1 shows the population levels of the major geographic areas between 1800 and the present.

3.2.3. Population Scenarios

3.2.3.1. Population Projections Used in Emission Scenarios

Since the IPCC was first convened in 1988, its Working Group III has generated two distinct series of emissions scenarios: the 1990 Scientific Assessment (SA90) series of four scenarios (Houghton et al., 1990), and the IS92 series of six scenarios (Houghton et al., 1992, 1995; Pepper et al., 1992). The four SA90 scenarios all used the same median population projection - the World Bank 1987 projection (Zachariah and Vu, 1988). The IS92a-f series made use of three different projection variants, the World Bank (World Bank, 1991) 1991 projection and the United Nations (UN, 1992) 1992 medium-high and medium-low projections.

Wexler (1996) surveyed the world population projections used in GHG emissions scenarios since 1990. Of the models surveyed, all but one employed the World Bank central projections. The sole exception - the DICE model of Nordhaus (1993; Nordhaus and Yohe, 1983) - used algorithmic projections based on assumed declining population growth rates. As noted in Chapter 2, many long-term emission scenarios available in the literature do not even report their underlying population projections. In general, World Bank projections have been more heavily employed than the UN projections, apparently because of the shorter time horizon and longer cycle time of the UN Long Range projections. UN Revisions until 1994 extended only to 2025 (now to 2050), which is too short for emissions scenarios. The UN Long Range series is revised less frequently, with a six-year interval between the previous two UN Long Range projections. The World Bank, in contrast, updates its published projection every two years, and it has always been a long-range projection, out to year 2150. In addition, the World Bank maintains a country-level disaggregation throughout its projection, unlike the UN Long Range series, which switches to a nine-region summary.

In 1994, however, the World Bank discontinued the publication of population projections, even if these continue to be generated for internal uses. In contrast to the past dominance of World Bank projections, the SRES scenario series instead employs published projections from the International Institute for Applied Systems Analysis (IIASA) along with the UN's Medium Long Range projection.

IPCC Homepage